1 paper · 1 filter
Kevin Ren, Manish Raghavan, Nikhil Garg
Deployed approaches for AI text detection often rely on training-time access to labeled datasets of both human-written and AI-generated text. This approach is vulnerable to three t…